The Effect of an Obesity Prevention Intervention Among Specific Subpopulations: A Heterogeneity of Treatment Effect Analysis of the Greenlight Trial

Abstract

Background:

Understanding how different populations respond to a childhood obesity intervention could help optimize personalized treatment strategies, especially with the goal to reduce disparities in obesity.

Methods:

We conducted a secondary analysis of the Greenlight Cluster Randomized Controlled Trial, a health communication focused pediatric obesity prevention trial, to evaluate for heterogeneity of treatment effect (HTE) by child biological sex, caregiver BMI, caregiver reported race and ethnicity, primary language, and health literacy. To examine HTE on BMI z-score from 2 to 24 months of age, we fit linear mixed effects models.

Results:

We analyzed 802 caregiver–child pairs, of which 52% of children were female, 58% of households reported annual family income of <$20,000, and 83% did not have a college degree. We observed evidence to suggest HTE by primary language (p = 0.047 for Spanish vs. English) and the combination of primary language and health literacy (p = 0.01). There was insufficient evidence to suggest that the Greenlight intervention effect differed by biological sex, caregiver BMI, or by race/ethnicity.

Conclusions:

This HTE analysis found that the Greenlight obesity prevention intervention had a more beneficial effect on child BMI z-score over 2 years for children of caregivers with limited health literacy and for caregivers for whom Spanish was the primary language.

Introduction

Disparities in the prevalence of childhood obesity by race, ethnicity, household income, caregiver BMI, and caregiver health literacy are evident by early childhood, such that children from lower socioeconomic strata, those with caregivers of higher BMI and lower health literacy, and those from racial and ethnic minority populations have the highest prevalence of childhood obesity.^1–4 Despite a general consensus that social determinants of health contribute to disparities in childhood obesity, little progress has been made in developing solutions to reduce those health disparities.^5,6

Developing a more complete understanding of how various populations respond to obesity interventions, particularly among populations with a disproportionate prevalence of obesity, would help inform the development of strategies to reduce health disparities in childhood obesity and its associated long-term consequences.

Behavioral interventions have an important role to play in the prevention of childhood obesity.^7–9 Recently published clinical practice guidelines by the American Academy of Pediatrics indicate the need for focused attention on childhood obesity prevention through the implementation of early interventions with a particular focus on reducing health disparities.¹⁰ Yet, few well-conducted trials of behavioral interventions are conducted among populations with low socioeconomic resources and among racial and ethnic minority groups.¹¹

In addition, trials that have been conducted among these populations for obesity prevention in both infancy and early childhood consistently fail to achieve similar effect sizes as those achieved in populations with high socioeconomic resources or in majority White groups.^12–19 Thus, there is a need to more carefully evaluate which behavioral interventions may be most effective in populations with low socioeconomic resources and among racial and ethnic minority groups, with the goal being to reduce health disparities in childhood obesity by providing tailored behavioral interventions.

The Greenlight intervention was an obesity prevention intervention delivered from 2 months to 2 years of age at routine preventive well-child care visits in pediatric primary care continuity clinics.²⁰ The intervention was specifically designed to support pediatric primary care providers in applying principles of low-health literacy and culturally appropriate communication while engaging caregivers in a shared goal-setting practice around healthy behaviors. Greenlight was tested in a cluster-randomized trial (N = 802 caregiver–child pairs) that observed improvements in child weight trajectories through the first 18 months of life, although did not have an effect on the primary or secondary outcomes at 24 months of life.²¹

The purpose of this analysis is to determine whether the intervention had differential effectiveness [i.e., exploratory heterogeneity of treatment effect (HTE) analysis] among certain subpopulations enrolled in the trial.²² Because the intervention was designed to support effective communication for caregivers with low health literacy and for parents from nonmajority cultures, we hypothesized that the intervention would be more effective (primary outcome: BMI z-score) for those population subgroups when compared with parents with adequate health literacy and those who self-identified as non-Hispanic White.

Methods

Study Design

We conducted a secondary analysis of the Greenlight Cluster Randomized Controlled Trial (NCT 01040897); the methods and results from the primary analysis have been previously published.^20,21 The Institutional Review Board at each of the participating medical centers approved the study. Written informed consent was obtained in the participant's preferred language (English or Spanish). The parent or legal guardian granted permission for participation. The study was cluster-randomized, with each of the four sites being randomly assigned to either the intervention or attention control condition. Because this was a cluster-randomized trial study staff were not blinded to group assignment during the study.

The study was performed at pediatric primary care clinics (i.e., continuity clinics) among pediatric residents at four academic medical centers: New York University, The University of North Carolina at Chapel Hill, The University of Miami, and Vanderbilt University Medical Center. Two medical centers delivered the active intervention (obesity prevention), whereas two medical centers delivered the attention control (injury prevention curriculum).

At the intervention sites, pediatric residents received training in health communication (including addressing health literacy, shared goal setting, language and cultural issues, and use of interpreter services), and were given Greenlight toolkits in English or Spanish to share with families as part of counseling along with shared goal setting at each scheduled preventive visit. Toolkits included low literacy age-specific Greenlight booklets that were specially designed using a health literacy-informed approach, with attention paid to literacy demand (sixth-grade reading level target), scope limited to key information centered on priority actions and behaviors, and extensive inclusion of photos and other visuals to support text information.

The toolkits were designed to support healthy behaviors throughout the first 2 years of life, including a focus on feeding behaviors, recognizing and addressing satiety cues, appropriate timing of introducing solids, portion sizes, diet quality, reducing screen time, limiting sugary beverages, and promoting appropriate physical activity. Families of young children—including parents with limited literacy and limited English-proficiency skills—participated in the codesign of the Greenlight toolkits.

Participants

Participants were recruited at the 2-month well-child check at their primary care provider's office between April 28, 2010 and July 24, 2012. Child eligibility criteria included age 6–12 weeks at the time of study enrollment. Parent eligibility criteria included being English- or Spanish-speaking and agreement to participate in the study for 2 years. Child exclusion criteria included being born before 32 weeks gestation, birth weight <1500 g, weight/length percentile less than the third percentile at 2 months of age, a diagnosis of failure to thrive, or a known medical problem that would impact child growth (e.g., metabolic disease or uncorrected congenital heart disease).

Parent exclusion criteria included significant mental or neurological illness likely to impair their ability to participate, age <18 years, or poor visual acuity. For the present analysis, an additional data eligibility criterion was used such that children were required to have at least one follow-up weight and length measure after the baseline measurement.

Outcomes

The primary outcome for the current analysis is child BMI z-score by World Health Organization (WHO) standards at 2 years, which was calculated from length and weight measures collected during routine clinical care by clinic staff who were trained in growth measurement. Weight and length measures were obtained from the electronic health record if they were taken on the same day as a part of a routine well-child check. These measures were collected at approximately 2, 4, 6, 9, 12, 15, 18, and 24 months, based on the usual timing of well-child checks.

In addition, caregivers completed a survey at the baseline timepoint, in English or Spanish based on their preference, which collected data on sociodemographic characteristics and specifically included single items to assess annual household income and participation (for mother and child) in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). Primary language was defined as primary reading language, as the intervention involved distribution of written booklets in the parents' preferred language (Spanish or English).

Caregiver health literacy was assessed by the Parental Health Literacy Activities Test (PHLAT), which has been validated in both English and Spanish.^23,24 The PHLAT is an 8-item scale that is scored by adding each of the correct items for a possible range of 0–8, with higher scores representing higher health literacy.

Statistical Analysis

We summarized baseline characteristics stratified by primary language, using percentiles for continuous variables and proportions for categorical variables. Among the 865 caregiver–child dyads randomized into the study, 802 had at least 1 follow-up visit and are included in analyses.

To examine HTEs on BMI z-score from 2 to 24 months of age, we fit linear mixed effects models with maximum likelihood using random intercepts and slopes (for age) to acknowledge within-child correlation (see Supplementary Appendix SA1 for full details). Based on the authors' knowledge of factors that have been previously associated with childhood obesity, we identified the following variables to test for HTE: child biological sex, caregiver BMI (≥30 kg/m² at baseline, <30 kg/m²), child race and ethnicity (Hispanic, non-Hispanic White, non-Hispanic Black, or non-Hispanic other race), caregiver's primary language, defined as preferred reading language (Spanish or English), and caregiver's health literacy (score on the PHLAT dichotomized at <6 vs. ≥6, which was based on the distribution of PHLAT in our study sample).

Because this was a cluster randomized clinical trial with only four clusters, potential confounding by site was a concern and so we built mixed models that included covariates as well as interactions among three key variables: site, age, and each heterogeneity variable (HV). For each value of the HV, we estimated intervention effects using the difference between the average estimated trajectories at the Greenlight and comparator sites.

We then examined heterogeneity in the trajectory differences across levels of the HV. Specifically, we used linear contrasts of the site-by-age interactions to ascertain the intervention effect across all ages by subtracting the covariate-adjusted average trajectory of the two control sites from the two intervention sites. We report intervention effects across follow-up and confidence intervals for distinct levels of the HV, and we use Wald statistics to test for HTE over time using a two-sided 0.05 significance level (see Supplementary Appendix SA1 for full details).

We included the following covariates as fixed effects in the linear mixed effect model: baseline child characteristics (sex, race/ethnicity, age, baseline BMI z-score, and site), baseline caregiver characteristics (family income, caregiver education, caregiver PHLAT, caregiver BMI, caregiver age, and WIC participation), and time-varying values (age, age by baseline BMI z-score interaction, and age by site interaction). For each HV analysis, we included in the model the HV, the HV by site interaction, the HV by time interaction, and the HV by site by time interaction.

Covariates age, caregiver BMI, and caregiver age were entered into models using flexible restricted cubic spline functionals, except when caregiver BMI was a HV, it was entered into the model as a dichotomous variable. Because enrollment in the non-Hispanic other race/ethnicity was low (N = 32) those patients were removed from analyses of HTE by race/ethnicity, but were included in all other analyses.

To address missing data, including baseline data and follow-up BMI, we conducted multiple imputation with chained equations using predictive mean matching.²⁵ The percent of missing data for each variable at baseline and for weight and length at each follow-up timepoint is included in Table A1 of Supplementary Appendix SA1. We generated 100 imputation data sets and fit the corresponding linear mixed effects models and associated summaries for each imputation. We then combined results across the imputations using Rubin's Rules.²⁶ All analyses were conducted by using R version 3.6.3.^27,28

Results

A total of 802 participants had at least one follow-up visit and are included in analyses. In this sample, 52% of children were female, 58% of households reported annual family income of <$20K, 85% of caregivers reported at least some WIC support at enrollment, and 83% did not have a college degree (see Table 1 for full demographic characteristics). Approximately 50% of caregivers were Hispanic, 35% of caregivers reported Spanish as their primary language, and 43% of caregivers reported limited English proficiency.

Table 1.

Demographic Characteristics of Children and Caregivers Included in Analyses by Caregiver Reading Language

Variables	N	Combined	English	Spanish
Variables	N	% (N = 802)	% (N = 520)	% (N = 281)
Child sex	802
Male		48 (388)	48 (248)	49 (139)
Female		52 (414)	52 (272)	51 (142)
Caregiver race and ethnicity	802
All other, non-Latino		4 (32)	6 (32)	0 (0)
Black, non-Latino		28 (224)	43 (223)	0 (0)
Latino		50 (402)	23 (121)	100 (281)
White, non-Latino		18 (144)	28 (144)	0 (0)
Annual household income	795
<$10,000		30 (239)	25 (130)	39 (109)
$10,000–19,999		28 (219)	23 (120)	35 (99)
$20,000–39,999		24 (190)	27 (139)	18 (51)
$40,000–59,999		8 (66)	11 (59)	2 (7)
≥$60,000		7 (58)	10 (52)	2 (6)
Caregiver did not know		3 (23)	3 (14)	3 (9)
Caregiver age, years	798	27 [23, 31]	26 [22, 31]	29 [25, 33]
Caregiver BMI, kg/m²	724	27 [24, 32]	28 [24, 33]	27 [24, 31]
PHLAT at month 2	789	5 [3, 6]	6 [4, 7]	4 [2, 6]
Birth weight, kg	797	3.3 [2.9, 3.6]	3.2 [2.9, 3.5]	3.4 [3.0, 3.7]
Caregiver born in United States	799
Yes		50 (398)	76 (394)	1 (4)
No		50 (401)	24 (123)	99 (277)
Caregiver English-proficiency level	799
Very well		57 (457)	86 (443)	5 (14)
Well		14 (109)	11 (58)	18 (50)
Not well		19 (151)	2 (11)	50 (140)
Not at all		10 (82)	1 (5)	27 (77)
Caregiver sex	802
Male		4 (32)	4 (23)	3 (8)
Female		96 (770)	96 (497)	97 (273)
Caregiver education	798
College		17 (139)	23 (120)	7 (19)
HS or some college		56 (447)	63 (327)	43 (120)
Less than HS		27 (212)	14 (70)	51 (142)
Enrolled in WIC program, mother or child	798
Any WIC		85 (678)	80 (415)	94 (263)
No WIC		15 (120)	20 (102)	6 (18)

Categorical variables are presented as percentages (counts) and continuous variables with median (25th and 75th percentiles).

HS, high school; PHLAT, Parental Health Literacy Activities Test (measure of health literacy); WIC, special supplemental nutrition program for women, infants and children.

The median PHLAT score was 5 with interquartile range 4–6, which implies that >50% of participants would be considered to have inadequate health literacy.^23,24 Those whose primary reading language was Spanish reported lower income, lower educational attainment, and somewhat higher WIC utilization rates. Finally, BMI measures were missing at 18.7% of potential visits in our study, with higher levels of missingness at later timepoints, including 39.7% missing BMI at the 24-month timepoint (See Table A1 in Supplementary Appendix SA1 for complete details on missing BMI).

Because of this, we conducted a comparison of children who missed at least one visit with weight and length/height (i.e., a calculated BMI) and those who had all scheduled measures, with no major differences observed (see Table A2 in Supplementary Appendix SA1).

Figure 1 shows results from the HTE analysis. It displays a series of adjusted BMI z-score differences across distinct level of each of the HVs. Note that primary study results showed adjusted BMI z-scores at the Greenlight sites to be lower than those at the control sites during much of the study period (through 18 months).²¹ In this study, we observed evidence to suggest HTE by primary reading language (p = 0.047 for Spanish vs. English) and the combination of low health literacy and primary reading language (p = 0.01).

Figure 1.

Heterogeneity of treatment effect. Each embedded figure shows the difference in BMI z-score between intervention and control groups for specific subgroups of interest. These differences are displayed for the first 2 years of life and a negative BMI z-score difference represents a desirable intervention effect. Model-based estimates are shown for specific subgroups of interest, with a corresponding p-value for the interaction between the variable of interest and the treatment effect. PHLAT, Parental Health Literacy Activities Test. Color image is available online.

Specifically, in those whose primary reading language was Spanish with PHLAT <6 we observed that at 6, 12, 18, and 24 months, children participating at the Greenlight sites has adjusted BMI z-scores that were −0.24 (−0.37, −0.11), −0.54 (−0.84, −0.25), −0.60 (−0.95, −0.24), and −0.38 (−0.77, 0.01) units lower, respectively, than those who participated at the control sites. This is in contrast with those whose primary reading language was English or PHLAT ≥6, where BMI z-score differences at 6, 12, 18, and 24 months between the Greenlight sites and control sites were −0.03 (−0.10, 0.04), −0.05 (−0.21, 0.11), 0.03 (−0.16, 0.22), and 0.22 (0.00, 0.44).

There is insufficient evidence to suggest HTE for the Greenlight intervention by biological sex, caregiver BMI, or by race/ethnicity. Full model outputs for each of these models are shown in the table in Supplementary Appendix SA1.

Discussion

In this secondary analysis of the Greenlight Randomized Trial, we found HTE for important population subgroups. The intervention was more effective at reducing child BMI z-score for the first 2 years of life for families with limited health literacy, for whom Spanish was the primary language, and even more so for families who were in both groups. Importantly, the Greenlight intervention was specifically designed for families with low health literacy and for whom English was not the primary language.

There was insufficient evidence to suggest HTE by a child's biological sex, caregiver BMI, or by race or ethnicity. These findings suggest that the Greenlight intervention was differentially effective for populations that have a disproportionate prevalence of childhood obesity. Future studies should consider whether behavioral obesity interventions, such as Greenlight, can specifically reduce health disparities in childhood obesity.

Compared with disparities in race and ethnicity, childhood obesity disparities related to parent health literacy have been less-well explored. In our analysis, a parent's health literacy appeared to contribute to modification of the intervention effect. Health literacy is defined as “the degree to which individuals have the ability to find, understand, and use information and services to inform health-related decisions and actions for themselves and others.”²⁹ One reason that interventions designed to improve early child weight trajectories may be less effective among low-income and minority populations is structural barriers faced by those with low health literacy.^30,31

Low parent health literacy has been consistently associated with the development of early childhood obesity in observational studies.³² In addition, low health literacy is more common among low-income and minority families.³³ Low parent health literacy may contribute to the development of childhood obesity by making it more difficult for parents to understand anticipatory guidance aimed at promoting healthy behaviors for obesity prevention.³⁴ In our own study, low parent health literacy was associated with health behaviors that contribute to obesity as early as infant age 2 months.³²

We may have observed greater effectiveness of the Greenlight intervention among parents with low health literacy because the booklet design incorporated characteristics we hypothesized would meet the needs of parents with low health literacy, including the use of many photos and visuals, and brief limited text. Developing intervention strategies that take advantage of effective communication strategies for parents with low health literacy may, therefore, be an effective strategy for reducing health disparities in early child obesity.

Similarly, a parent's preferred language was a statistically significant effect modifier of the intervention effect. Previous studies have suggested that caregivers who speak Spanish primarily are less commonly satisfied with physician communication in the US health care system.^35–37 Our own study also showed disparities in physician communication satisfaction that was associated with language but not ethnicity.³⁸ The results from this study suggest that using tools to facilitate culturally sensitive and appropriate communication between pediatric primary care providers and patients for whom English is not the primary language may be one way to facilitate more effective management of obesity prevention in primary care settings.

We observed heterogeneity of effect based on primary language and health literacy level but not for income, race, or ethnicity. The reasons for why some of these populations experienced differential intervention effectiveness are unclear. One potential explanation is that the Greenlight toolkit was specifically designed to support Spanish-speaking families through cultural and linguistic tailoring and also adhered to literacy-sensitive principles (white space, color-coding, multiple pictures, and infographics, using short simple sentences).

This analysis had several limitations. First, the study was a cluster randomized trial with only four sites, which could limit the power to detect important subgroup differences in the intervention effect. Moreover, there may be imbalances in characteristics between participants of the intervention and control arms given the small number of sites randomized, but we controlled for a number of baseline covariates to account for site to site differences. At the same time, it could also increase type 1 error rates due to chance differences between the two Greenlight sites and two comparator sites not attributable to the intervention.

It should also be noted that the analysis plan was exploratory (i.e., not prespecified), which should limit the implications of the current findings as primarily hypothesis generating, and not definitive evidence that the intervention was more effective in one population or another. However, we tested differences in treatment effect among the subgroups for whom the Greenlight intervention was designed. Furthermore, nonprespecified analyses can be instructive to determine important receiver characteristics for large trials.^39,40

Finally, although 18.7% of follow-up visits were missed, an additional 5.3% of visits did not collect both weight and length (i.e., 24% of follow-up visits did not have BMI z-score) and the differences between observed and missing visits could explain observed associations. It should be noted that multiple imputation relies on data being missing at random.

To the extent there is missingness (e.g., 40% of visits at 15 and 24 months) and the missing at random assumption is incorrect, results could be biased. Note that Table A2 in Supplementary Appendix SA1 shows characteristics of those with no missing follow-up weights and lengths to those with at least some missing data. Finally, as an exploratory analysis, this study may have been underpowered to detect three-way interactions, which may have led to type 2 error.

In conclusion, this exploratory HTE analysis suggested that the Greenlight obesity prevention intervention had a stronger effect on early childhood BMI z-score trajectory for children of caregivers with limited health literacy and for caregivers for whom Spanish was the primary language. This suggests that future research should focus on understanding how obesity prevention interventions impact specific population subgroups, with the goal of developing strategies to reduce health disparities in childhood obesity.

Impact Statement

Future research should focus on understanding how obesity prevention interventions impact not just overall populations but specific population subgroups to inform the development of best strategies to reduce disparities in childhood obesity.

Footnotes

Acknowledgments

Preliminary data from this analysis were presented at The Obesity Society Annual Meeting in 2023, as a poster presentation. The previously presented abstract is available online, at Obesity (Silver Spring), 31:1–4.

Authors' Contributions

All authors were responsible for the conceptualization of the analysis and the review and editing of the article. W.J.H. was responsible for drafting the original article. J.S.S. and A.B. were responsible for overseeing the data curation and implementing the analysis plan, with input from all the authors. H.S.Y., R.L.R., K.B.F., A.M.D., L.S., and E.M.P. were responsible for initial funding acquisition.

Funding Information

All phases of this study were supported by the Eunice Kennedy Shriver Institute for Child Health and Development, NICHD (Grant No. R01HD049794), with supplemental funding from CDC and OBSSR (Grant Nos. R01HD059794-04S1 and R01HD059794-26 04S2). Parts of the study were supported by the National Institutes of Health's National Center for Advancing Translational Sciences through its Clinical and Translational Science Awards Program (CTSA), grant nos. 1UL1RR029893 (NYU), UL1TR000445 (Vanderbilt University), and UL1RR025747 (UNC-Chapel Hill). During the time the study was conducted, H.S.Y. was supported by a grant under the Robert Wood Johnson Foundation Physician Faculty Scholars Program and HRSA (12-191-1077-Academic Administrative Units in Primary Care) and by funding from the KiDS of NYU Langone Foundation.

Author Disclosure Statement

The authors have no conflicts of interest relevant to this article to disclose.

Supplementary Material

References

Isong

, Rao

, Bind

M-A

, et al. Racial and ethnic disparities in early childhood obesity. Pediatrics, 2018; 141(1):e20170865.

, Staiano

. Trends in obesity prevalence among children and adolescents aged 2 to 19 years in the US from 2011 to 2020. JAMA Pediatrics, 2022; 176(10):1037–1039.

Chari

, Warsh

, Ketterer

, et al. Association between health literacy and child and adolescent obesity. Patient Ed Counsel, 2014; 94(1):61–66.

Min

, Xue

, Wang

. Association between household poverty dynamics and childhood overweight risk and health behaviours in the United States: A 8-year nationally representative longitudinal study of 16 800 children. Pediatr Obes, 2018; 13(10):590–597.

Min

, Goodale

, Xue

, et al. Racial-ethnic disparities in obesity and biological, behavioral, and sociocultural influences in the United States: A systematic review. Adv Nutrit, 2021; 12(4):1137–1148.

Huang

, Drewnosksi

, Kumanyika

, Glass

. A systems-oriented multilevel framework for addressing obesity in the 21st century. Prevent Chronic Dis, 2009; 6(3):A82.

Brown

, Moore

THM

, Hooper

, et al. Interventions for preventing obesity in children. Cochrane Database Syst Rev, 2019; 7:CD001871.

O'Connor

, Evans

, Burda

, et al. Screening for obesity and intervention for weight management in children and adolescents: Evidence report and systematic review for the US Preventive Services Task Force. JAMA, 2017; 317(23):2427–2444.

Colquitt

, Loveman

, O'Malley

, et al. Diet, physical activity, and behavioural interventions for the treatment of overweight or obesity in preschool children up to the age of 6 years. Cochrane Database Syst Rev, 2016; 3:CD012105.

10.

Hampl

, Hassink

, Skinner

, et al. Clinical practice guideline for the evaluation and treatment of children and adolescents with obesity. Pediatrics, 2023; 151(2):e2022060640.

11.

Narzisi

, Simons

. Interventions that prevent or reduce obesity in children from birth to five years of age: A systematic review. J Child Health Care, 2020; 25(2):320–334.

12.

Barkin

, Heerman

, Sommer

, et al. Effect of a behavioral intervention for underserved preschool-age children on change in body mass index: A randomized clinical trial. JAMA, 2018; 320(5):450–460.

13.

French

, Sherwood

, Veblen-Mortenson

, et al. Multicomponent obesity prevention intervention in low-income preschoolers: Primary and subgroup analyses of the NET-Works Randomized Clinical Trial, 2012–2017. Am J Public Health, 2018; 108(12):1695–1706.

14.

Robinson

, Matheson

, Wilson

, et al. A community-based, multi-level, multi-setting, multi-component intervention to reduce weight gain among low socioeconomic status Latinx children with overweight or obesity: The Stanford GOALS randomised controlled trial. Lancet Diabetes Endocrinol, 2021; 9(6):336–349.

15.

Robinson

, Matheson

, Kraemer

, et al. A randomized controlled trial of culturally tailored dance and reducing screen time to prevent weight gain in low-income African American girls: Stanford GEMS. Arch Pediatr Adolesc Med, 2010; 164(11):995–1004.

16.

Dietz

. We need a new approach to prevent obesity in low-income minority populations. Pediatrics, 2019; 143(6):e20190839.

17.

Butler ÉM, Fangupo

, Cutfield

, et al. Systematic review of randomised controlled trials to improve dietary intake for the prevention of obesity in infants aged 0–24 months. Obes Rev, 2021; 22(2):e13110.

18.

Brown

, Moodie

, Sultana

, et al. A scoping review of outcomes commonly reported in obesity prevention interventions aiming to improve obesity-related health behaviors in children to age 5 years. Obes Rev, 2022; 23(5):e13427.

19.

Larsen

, Karssen

, van der Veek

. Targeting food parenting practices to prevent early child obesity risk requires a different approach in families with a lower socioeconomic position. Front Public Health, 2022; 10:1012509.

20.

Sanders

, Perrin

, Yin

, et al. “Greenlight study”: A controlled trial of low-literacy, early childhood obesity prevention. Pediatrics, 2014; 133(6):e1724–e1737.

21.

Sanders

, Perrin

, Yin

, et al. A health-literacy intervention for early childhood obesity prevention: A cluster-randomized controlled trial. Pediatrics, 2021; 147(5):e2020049866.

22.

Varadhan

, Segal

, Boyd

, et al. A framework for the analysis of heterogeneity of treatment effect in patient-centered outcomes research. J Clin Epidemiol, 2013; 66(8):818–825.

23.

Kumar

, Sanders

, Perrin

, et al. Parental understanding of infant health information: health literacy, numeracy, and the Parental Health Literacy Activities Test (PHLAT). Acad Pediatr, 2010; 10(5):309–316.

24.

Yin

, Sanders

, Rothman

, et al. Assessment of health literacy and numeracy among Spanish-Speaking parents of young children: Validation of the Spanish Parental Health Literacy Activities Test (PHLAT Spanish). Acad Pediatr, 2012; 12(1):68–74.

25.

Raghunathan

, Lepkowski

, Van Hoewyk

, et al. A multivariate technique for multiply imputing missing values using a sequence of regression models. Surv Methodol, 2001; 27:85–95.

26.

Little

RJA

, Rubin

. Statistical Analysis with Missing Data, 2nd ed. Wiley: Hoboken, NJ, USA; 2002.

27.

R: A Language and Environment for Statistical Computing [Computer Program]. R Foundation for Statistical Computing: Vienna, Austria; 2021.

28.

Harrell

. rms: Regression Modeling Strategies. R package version 6.1-0. 2020. Available from: https://CRAM.R-project.org/package=rms Accessed March 1, 2024.

29.

Santana

, Brach

, Harris

, et al. Updating health literacy for healthy people 2030: Defining its importance for a new decade in public health. J Public Health Manag Pract, 2021; 27(Suppl. 6):S258–S264.

30.

Williams

, Davis

, Parker

, et al. The role of health literacy in patient-physician communication. Fam Med, 2002; 34(5):383–389.

31.

White

, Thompson

, Rothman

, et al. A health literate approach to the prevention of childhood overweight and obesity. Patient Ed Counsel, 2013; 93(3):612–618.

32.

Yin

, Sanders

, Rothman

, et al. Parent health literacy and “obesogenic” feeding and physical activity-related infant care behaviors. J Pediatr, 2014; 164(3):577.e1–583.e1.

33.

Rikard

, Thompson

, McKinney

, et al. Examining health literacy disparities in the United States: A third look at the National Assessment of Adult Literacy (NAAL). BMC Public Health, 2016; 16:1–11.

34.

Marks

. Childhood obesity and parental health literacy. Adv Obes Weight Manag Control, 2015; 3(3):191–195.

35.

Halfon

, Inkelas

, Mistry

, et al. Satisfaction with health care for young children. Pediatrics, 2004; 113(Suppl. 5):1965–1972.

36.

Seid

, Stevens

, Varni

. Parents' perceptions of pediatric primary care quality: Effects of race/ethnicity, language, and access. Health Serv Res, 2003; 38(4):1009–1032.

37.

Steinberg

, Valenzuela-Araujo

, Zickafoose

, et al. The “battle” of managing language barriers in health care. Clin Pediatr, 2016; 55(14):1318–1327.

38.

Flower

, Skinner

, Yin

, et al. Satisfaction with communication in primary care for Spanish-speaking and English-speaking parents. Acad Pediatr, 2017; 17(4):416–423.

39.

Kent

, Hayward

. Limitations of applying summary results of clinical trials to individual patients: The need for risk stratification. JAMA, 2007; 298(10):1209–1212.

40.

Hayward

, Kent

, Vijan

, et al. Reporting clinical trial results to inform providers, payers, and consumers. Health Affairs, 2005; 24(6):1571–1581.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB